Effect of bypassing Rk on a Cathodyne PI

[Snicksound]

In trying to make this low power project a bit more aggressive, I was trying to add a little Presence boost. But since it's a cathodyne PI with the negative feedback applied at its input, mixed with the normal signal through resistors, the usual approach doesn't work. But for fun, I tried bypassing Rk (the 56k resistor to ground, aka R22 in the schematic below) with a 1nF cap and... it kinda does the job. But now I'm quite confused as to what is happening exactly.

Becke PI.png

Given that the NFB (at the bottom of the image) goes first through a 1M resistor, and then another 470k from to the junction of Rb and Rk, I don't think bypassing the extra 56k resistor does much to it, so there's gotta be something more to it.

Maybe I'm just getting a boost in highs by having a cap bypassing part of the resistance between the cathode and ground, but how does that affect the operation of the PI?

[thetragichero]

I'd imagine you're getting a crazy unbalanced output from that pi

[jcool666]

If i were to attempt this, I would try applying the nfb to the cathode of the gain stage immediately proceeding the PI. That seems to be the most straightforward way to me. Have you checked what the signal looks like on the grids of your output tubes? If you dont have a scope, check the AC voltage.

[Phil_S]

If i were to attempt this, I would try applying the nfb to the cathode of the gain stage immediately proceeding the PI. That seems to be the most straightforward way to me. Have you checked what the signal looks like on the grids of your output tubes? If you dont have a scope, check the AC voltage.

What Joe says. FWIW, as best I have gathered, the cathodyne/concertina/split load PI is not a "voltage device." It is a current device. It has a gain of 0. It puts out the same amount of current on each side. Because of this, it is a poor candidate for injecting NFB.

If you really want to know about it, here is what Merlin has to say:
https://www.google.com/url?sa=t&rct=j&q ... L445E93gul

[Snicksound]

If i were to attempt this, I would try applying the nfb to the cathode of the gain stage immediately proceeding the PI. That seems to be the most straightforward way to me. Have you checked what the signal looks like on the grids of your output tubes? If you dont have a scope, check the AC voltage.

The stage before that one is a DC cathode follower feeding the tone stack. Injecting NFB before the TS would be a little counter productive.

I copied this approach on the Marshall JTM1. It works well enough because it sounds much better than without the NFB. Plus the Master Volume just before it is amazingly transparent (sounds good all the way down to whisper levels)

It just makes it hard to implement a Presence control.

I was mostly curious of what was happening when bypassing the 56k. I can't wrap my head around it

[pdf64]

...I was mostly curious of what was happening when bypassing the 56k. I can't wrap my head around it

I think that (partially) bypassing the cathode of a cathodyne acts to change it into a regular common cathode gain stage at high frequencies, albeit one whose cathode is elevated above 0V.
So the (inverting) gain at the anode will be ca 1 at low frequencies and at high freqs -

from Aiken http://www.aikenamps.com/index.php/desi ... amplifiers
Av = (mu*Ra)/(Ra+ra)
= (100 x 56k//220k) / (56k//220k + 62k5)
= 42
but the output from the cathode will be 0, so the output stage will operate in SE rather than p-p mode.
So even though the gm doubling effect (in the class A area of operation) will be lost, the overall system gain will still be significantly higher.
But the power output will be much lower (all previous is with respect to low freq).

I was trying to add a little Presence boost. But since it's a cathodyne PI with the negative feedback applied at its input, mixed with the normal signal through resistors, the usual approach doesn't work

I don't think it works because the open loop gain is too low; for a presence boost to operate, the degree of NFB at low freq must be sufficient to reduce the gain; otherwise open loop operation at high freqs (due to the reduced degree of NFB) can't (by comparison) actually increase the effective system gain.

It's the same scenario as a Type 1/2/3 master vol (see Trainwreck Pages p41+) adjust to a low setting - a presence control can't work because the open loop gain is too low https://robrobinette.com/The_Trainwreck_Pages.htm

Hence I don't think it's feasible to add this type of presence control to your design.

Note that many amps have a presence control that operates a simple passive treble cut, eg red channel https://el34world.com/charts/Schematics ... tifier.pdf

Hope that makes sense

[Snicksound]

I think that (partially) bypassing the cathode of a cathodyne acts to change it into a regular common cathode gain stage at high frequencies, albeit one whose cathode is elevated above 0V.
So the (inverting) gain at the anode will be ca 1 at low frequencies and at high freqs -

from Aiken http://www.aikenamps.com/index.php/desi ... amplifiers
Av = (mu*Ra)/(Ra+ra)
= (100 x 56k//220k) / (56k//220k + 62k5)
= 42
but the output from the cathode will be 0, so the output stage will operate in SE rather than p-p mode.
So even though the gm doubling effect (in the class A area of operation) will be lost, the overall system gain will still be significantly higher.
But the power output will be much lower (all previous is with respect to low freq).

That's quite interesting. In a way I can see it now, if the 56k tail resistor is bypassed for AC, then there isn't much signal to feed into the power amp.
Tempted to try it with a bigger cap for a full bypass, out of curiosity. Not that I need this amp to be any less powerful (it's just an ECC99, it doesn't sound particularly good when cranked, and the MV works surprisingly well), but I like experimenting.

So far I didn't actually implement said "presence boost", I was just clipping stuff to try variations. But I have an unused switch position if I want to do it.

I don't think it works because the open loop gain is too low; for a presence boost to operate, the degree of NFB at low freq must be sufficient to reduce the gain; otherwise open loop operation at high freqs (due to the reduced degree of NFB) can't (by comparison) actually increase the effective system gain.

It's the same scenario as a Type 1/2/3 master vol (see Trainwreck Pages p41+) adjust to a low setting - a presence control can't work because the open loop gain is too low https://robrobinette.com/The_Trainwreck_Pages.htm

Hence I don't think it's feasible to add this type of presence control to your design.

Note that many amps have a presence control that operates a simple passive treble cut, eg red channel https://el34world.com/charts/Schematics ... tifier.pdf

Yeah I've pretty much abandoned the idea of a Marshall style Presence control, as did Marshall themselves on the JTM1. The Dual Rec uses a passive treble cut in part because it has no NFB, but this build doesn't sound to me that good without NFB (even though the preamp topology is not far from a Dual Rec). Then again, that's partly cause I use NFB for a fixed resonance boost (by way of the smallish 1000pF DC blocking cap in the NFB path) so try and make it sound a bit beefier (single ECC99, 10" speaker, small open back enclosure... needs a bit of help).

[roberto]

Maybe I'm just getting a boost in highs by having a cap bypassing part of the resistance between the cathode and ground, but how does that affect the operation of the PI?

Hi, you are doing the opposite: reducing the highs on the low non inverting side of the PI. Bypassing that 56k to ground (same applies for the upper one), is like adding a snubber on the plate resistor of a common gain stage.

Add a small capacitor with a reostat in series to the cathode of one of the previous stages, instead.